syndicalt/zaxy
by Various
Zaxy turns agent work into durable, auditable memory: a hash-chained Eventloom log as the source of truth, an embedded temporal knowledge graph for reasoning (local-first, no sidec
MCP
syndicalt/zaxy
Added 15 June 2026
Overview
Zaxy provides a durable, auditable memory system for AI agents using a hash-chained Eventloom log as the source of truth. It includes an embedded temporal knowledge graph for local-first reasoning, optional Neo4j or Postgres, and cited Memory Checkout for compact context. Model-facing MCP tools enable retrieval, capture, and feedback.
Best for
Best for
Developers building agent systems that need durable, auditable memory with local-first reasoning
Use cases
- Log agent actions in an immutable, tamper-evident event chain
- Query a temporal knowledge graph to support agent reasoning and decision-making
- Retrieve compact, cited context for LLM prompts via Memory Checkout
How to use
Install
pipx install zaxy-memory Tools exposed
NEO4J_PASSWORDMCP_ADMIN_TOKENPATHLIGHT_ACCESS_TOKEN
Tested with
Codex, Claude Code, Cursor, VS Code, Hermes Agent, LangGraph, CrewAI, AutoGen
Notes
Zaxy provides a durable, auditable memory system for AI agents using a hash-chained Eventloom log as the source of truth. It includes an embedded temporal knowledge graph for local-first reasoning, optional Neo4j or Postgres, and cited Memory Checkout for compact context. Model-facing MCP tools enable retrieval, capture, and feedback.
10 stars on GitHub. Last updated 2026-06-15. Licensed MIT.
Use cases
- Log agent actions in an immutable, tamper-evident event chain
- Query a temporal knowledge graph to support agent reasoning and decision-making
- Retrieve compact, cited context for LLM prompts via Memory Checkout
Pros
- Hash-chained log ensures audibility and integrity of agent memory
- Local-first temporal knowledge graph eliminates need for external databases by default
- Cited Memory Checkout reduces context footprint for efficient LLM calls
Cons
- Requires Python environment and MCP protocol integration
- Project is early-stage with limited community adoption (10 stars)
- Optional external database setup (Neo4j/Postgres) may add complexity for advanced use
Indexed from awesome-mcp-servers-punkpeye and enriched against its public facts.
Pros
- Hash-chained log ensures audibility and integrity of agent memory
- Local-first temporal knowledge graph eliminates need for external databases by default
- Cited Memory Checkout reduces context footprint for efficient LLM calls
Cons
- Requires Python environment and MCP protocol integration
- Project is early-stage with limited community adoption (10 stars)
- Optional external database setup (Neo4j/Postgres) may add complexity for advanced use
Pairs with
Other entries in the index that connect to this one. Click through to see the chain.
Get the free Developer’s Field Guide
A 27-page field guide to the AI coding workflow with Claude. Claude Code, MCP servers, the prompt patterns that work, and what to delegate. Free.
Enter your work email. We send it straight over, plus a few short notes worth knowing. Unsubscribe any time.